3D genome of multiple myeloma reveals spatial genome disorganization associated with copy number variations

Hi-C technique has been widely applied to study the three-dimensional architecture of the whole genome. Genome structures such as compartment A/B, TAD (topologically associated domain) and chromatin loops can be identified from Hi-C data in both normal cells of human and other species, and are found to be associated with features such as epigenetic markers, DNA-binding proteins and gene expression. But such technique had been rarely used in cancer studies. Here we used Hi-C to study the aneuploid cancer genomic architecture in multiple myeloma cells. Our results indicate that Hi-C interaction matrix of cancer cells is affected by CNVs and should be adjusted for copy number. After correcting this CNV bias, we found a significant overlapping between the boundaries of CNV blocks and boundaries of TADs, which suggests that TAD boundaries are fragile sites for CNV breakpoints. In addition, the compartment A/B switching is associated with differential gene expression, from which we found important genes that are related with multiple myeloma. We build a 3D structure model of the aneuploidy genome and found that there are great changes both in the whole genome spatial interactome and local chromosome territories. In summary, our research builds the first 3D genome interaction maps of multiple myeloma and the first time notice this CNV-driven bias in Hi-C studies, which may deepen our understanding of changes in cancer 3D genome. Overall design: Integrated analysis of 3D genome architecture, gene expression, and genome structure variations in multiple myelomas.